TODO.md

Urgent

• issue when there is no replacement list
• decouple keywords match and adding tags
• There is bug in terms of the implementation of fuzz.partial_ratio(), the result of the score is not consistent, maybe related to this issue. (It seems both partial_ratio and token_set_ration are not enough, one is too strict one is too weird).

Near term

• tests by pytest
• travis CI
• update README
• setup doc by sphnix
• add CHANGELOG
• unfied subject key of Paperls
• auto classify the announce date of papers
• link of authors

Furture plan

• paper metadata into database
• auto generate paper-style text
• webapp for arxiv analysis
• more machine learning techinques on arxiv papers to extract hot trend
• paper relevance and recommendations